Benefit Segmentation:

A Dedsion-oriented

Research ToolRUSSELL I. HALEY

According to this ar+icle,mos+ techniques of marketsegmentation rely only onDESCRIPTIVE factors pertain-ing to purchasers and arenot efficient predictors offuture buyer behavior. Theauthor proposes an ap-proach whereby market seg-ments are delineated firston the basis of factors witha CAUSAL relationship tofuture purchase behavior.The belief underlying thissegmentat ion strategy isthat the benefits which peo-ple are seeking in consuminga given product are thebasic reasons for the exist-ence of true market seg-ments.

Journal of Marketing, Vol. 32 (July.1968), pp. 30-35.

M ARKET segmentation has been steadily moving toward centerstage as a topic of discussion in marketing and research

circles. Hardly a conference passes without at least one sessiondevoted to it. Moreover, in March the American ManagementAssociation held a three-day conference entirely concerned withvarious aspects of the segmentation problem.

According to Wendell Smith. "Segmentation l̂ based upon de-velopments on the demand side of the market and represents arational and more precise adjustment of product and marketingeffort to consumer or user requirements."^ The idea that all mar-kets can be profitably segmented has now received almost as wide-spread acceptance as the marketing concept itself. However, prob-lems remain. In the extreme, a marketer can divide up his marketin as many ways as he can describe his prospects. If he wishes,he can define a left-handed segment, or a blue-eyed segment, or aGerman-speaking segment. Consequently, current discussion re-volves largely around which of the virtually limitless alternativesis likely to be most productive.

Segmentation MethodsSeveral varieties of market segmentation have been popular in

the recent past. At least three kinds have achieved some degreeof prominence. Historically, perhaps the first type to exist wasgeographic segmentation. Small manufacturers who wished tolimit their investments, or whose distribution channels were notlarge enough to cover the entire country, segmented the U. S.market, in effect, by selling their products only in certain areas.

However, as more and more brands became national, the sec-ond major system of segmentation—demographic segmentation—be-came popular. Under this philosophy targets were defined asyounger people, men, or families with children. Unfortunately, anumber of recent studies have shown that demographic variablessuch as age, sex, income, occupation and race are, in general, poorpredictors of behavior and, consequently, less than optimum basesfor segmentation strategies.-

1 Wendell R. Smith, "Product Differentiation and Market Segmentationas Alternative Product Strategies," JOURNAL OF MARKETING. Vol. XXI(July, 1956). pp. 3-8.

- Ronald E. Frank, "Correlates of Buying Behavior for Grocery Prod-ucts," JOURNAL OF MARKETING. Vol. 31 (October, 1967).pp. 48-53; Ron-ald E. Frank. William Massy, and Harper W. Boyd. Jr., "Correlatesof Grocery Product Consumption Rates," Journal of Marketing Re-search, Vol. 4 (May. 1967), pp. 184-190; and Clark Wilson, "Home-maker Living Patterns and Marketplace Behavior—A PsychometricApproach." in John S. Wright and Jac L. Goldstucker, Editors. NewIdeas for Successful Marketing, Proceedings of 1966 World Congress(Chicago: American Marketing Association, June, 1966), pp. 305-331.

30

Benefit Segmentation: A Decision-oriented Research Tool 31

More recently, a third type of segmentation hascome into increasing favor—volume seginentation.The so-called "heavy half" theory, popularized byDik Twedt of the Oscar Mayer Company,-' pointsout that in most product categories one-half of theconsumers account for around 80'^ of the consump-tion. If this is true, the argument goes, shouldn'tknowledgeable marketers concentrate their effortson these high-volume consumers? Certainly theyare the most valuable consumers.

The trouble with this line of reasoning is that notall heavy consumers are usually available to the samebrand—because they are not all seeking the samekinds of benefits from a product. For example,heavy coffee drinkers consist of two types of con-sumers—those who drink chain store brands andthose who drink premium brands. The chain storecustomers feel that all coffees are basically alike and,because they drink so much coffee, they feel it issensible to buy a relatively inexpensive brand. Thepremium brand buyers, on the other hand, feel thatthe few added pennies which coffees like Yuban.Martinson's, Chock Full O'Nuts, and Savarin costare more than justified by their fuller taste. Obvi-ously, these two groups of people, although they areboth members of the "heavy half" segment, are notequally good prospects for any one brand, nor canthey be expected to respond to the same advertisingclaims.

These three systems of segmentation have beenused because they provide helpful guidance in theuse of certain marketing tools. For example, geo-graphic segmentation, because it describes the mar-ket in a discrete way, provides definite direction inmedia purchases. Spot TV, spot radio, and news-papers can be bought for the geographical segmentselected for concentrated effort. Similarly, demo-graphic segmentation allows media to be boughtmore efficiently since demographic data on readers,viewers, and listeners are readily available for mostmedia vehicles. Also, in some product categoriesdemographic variables are extremely helpful in dif-ferentiating users from non-users, although they ai*etypically less helpful in distinguishing between theusers of various brands. The heavy-half philoso-phy is especially effective in directing dollars towardthe most important parts of the market.

However, each of these three systems of segmen-tation is handicapped by an underlying disadvantageinherent in its nature. All are based on an ex-postfacto analysis of the kinds of people who make upvarious segments of a market. They rely on de-scriptive factors rather than causal factors. Forthis reason they are not efficient predictors of future

buying behavior, and it is future buying behaviorthat is of central interest to marketers.

Benefit SegmentationAn approach to market segmentation whereby it

is possible to identify market segments by causalfactors rather than descriptive factors, might becalled "benefit segmentation." The belief under-lying this segmentation strategy is that the bene-fits which people are seeking in consuming a givenproduct are the basic reasons for the existence oftrue market segments. Experience with this ap-proach has shown that benefits sought by consumersdetermine their behavior much more accuratelythan do demogi-aphic characteristics or volume ofconsumption.

This does not mean that the kinds of datagathered in more traditional types of segmentationare not useful. Once people have been classifiedinto segments in accordance with the benefits theyare seeking, each segment is contrasted with all ofthe other segments in terms of its demography, itsvolume of consumption, its brand perceptions, itsmedia habits, its personality and life-style, and soforth. In this way, a reasonably deep understandingof the people who make up each segment can beobtained. And by capitalizing on this understand-ing, it is possible to reach them, to talk to them intheir own terms, and to present a product in themost favorable light possible.

The benefit segmentation approach is not new. Ithas been employed by a number of America's largestcorporations since it was introduced in 1961.•• How-ever, case histories have been notably absent fromthe literature because most studies have been con-tracted for privately, and have been treated confiden-tially.

3 Dik Warren Twedt, "Some Practical Applications ofthe 'Heavy Half Theory" (New York: AdvertisingResearch Foundation 10th Annual Conference, Oc-tober 6, 1964).

Russell I. Haley, "Experimental Research on Atti-tudes Toward Shampoos," an unpublished paper (Feb-ruary, 1961).

• ABOUT THE AUTHOR. Russell I.Haley is Vice President and CorporateResearch Director of D'Arcy Advertis-ing in New York City. Prior to his cur-rent position he was Vice President andAssociate Director of the Marketing andResearch Department at Grey Advertis-ing. While there, he developed newmethods for measuring attitudes, aunique way of segmenting markets byattitude patterns, and improved methodsof conducting large-scale market tests. Mr. Haley received hisM.B.A. from Columbia in marketing and statistics.

Mr. Haley is a past president of the Cleveland Chapter ofthe American Statistical Association. He is a member of theAmerican Marketing Association, the American Association forPublic Opinion Research, and the Executive Committee of theCopy Research Council. Mr. Haley is currently Chairman of anA.R.F, Committee dealing with attitude measurement and isteaching at Rutgers University.

32 Journal of Marketing, July, 1968

The benefit segmentation approach is based uponbeing able to measure consumer value systems indetail, together with what the consumer thinks aboutvarious brands in tbe product category of interest.While this concept seems simple enough, operation-ally it is very complex. There is no simple straight-forward way of handling the volumes of data thathave to be generated. Computers and sophisticatedmultivariate attitude measurement techniques are anecessity.

Several alternative statistical approaches can beemployed, among them the so-called "Q" techniqueof factor analysis, multi-dimensional scaling, andother distance measures.'"' All of these methods re-late the ratings of each respondent to those of everyother respondent and then seek clusters of individ-uals with similar rating patterns. If the itemsrated are potential consumer benefits, the clustersthat emerge will be groups of people who attachsimilar degrees of importance to the various benefits.Whatever the statistical approach selected, the endresult of the analysis is likely to be between threeand seven consumer segments, each representing apotentially productive focal point for marketingefForts.

Each segment is identified by the benefits it isseeking. However, it is the total configuration ofthe benefits sought which differentiates one segmentfrom another, rather than the fact that one segmentis seeking one particular benefit and another a quitedifl'erent benefit. Individual benefits are likely tohave appeal for several segments. In fact, the re-search that has been done thus far suggests thatmost people would like as many benefits as possible.However, the relative importance they attach to in-dividual benefits can differ importantly and, accord-ingly, can be used as an effective lever in segment-ing markets.

Of course, it is possible to determine benefit seg-ments intuitively as well as with computers andsophisticated research methods. The kinds of bril-liant insights which produced the Mustang and thefirst 100-millimeter cigarette have a good chanceof succeeding whenever marketers are able to tapan existing benefit segment.

However, intuition can be very expensive when itis mistaken. Marketing history is replete with ex-amples of products which someone felt could notmiss. Over the longer term, systematic benefit seg-mentation research is likely to have a higher pro-portion of successes.

But is benefit segmentation practical? And is ittruly operational? The answer to both of thesequestions is "yes." In effect, the crux of the prob-

Ronald E. Frank and Paul E. Green, "NumericalTaxonomy in Marketing Analysis: A Review Article,"Journal of Marketing Research, Vol. V (February,1968), pp. 83-98.

lem of choosing the best segmentation system is todetermine which has the greatest number of prac-tical marketing implications. An example shouldshow that benefit segmentation has a much widerrange of implications than alternative forms ofsegmentation.

An Example of Benefit SegmentationWhile the material presented here is purely illus-

trative to protect the competitive edge of companieswho have invested in studies of this kind, it is basedon actual segmentation studies. Consequently, it isquite typical of the kinds of things which are nor-mally learned in the course of a benefit segmenta-tion study.

The toothpaste market has been chosen as an ex-ample because it is one with which everyone isfamiliar. Let us assume that a benefit segmentationstudy has been done and four major segments havebeen identified—one particularly concerned with de-cay prevention, one with brightness of teeth, onewith the fiavor and appearance of the product, andone with price. A relatively large amount of sup-plementary information has also been gathered(Table 1) about the people in each of these segments.

The decay prevention segment, it has been found,contains a disproportionately large number of fam-ilies with children. They are seriously concernedabout the possibility of cavities and show a definitepreference for fluoride toothpaste. This is rein-forced by their personalities. They tend to be alittle hypochondriacal and, in their life-styles, theyare less socially-oriented than some of the othergi'oups. This segment has been named The Worriers.

The second segment, comprised of people who showconcern for the brightness of their teeth, is quitedifferent. It includes a relatively large group ofyoung marrieds. They smoke more than average.This is where the swingers are. Tbey are stronglysocial and their life-style patterns are very active.This is probably the group to which toothpastes suchas Macleans or Plus White or Ultra Brite wouldappeal. This segment has been named The Sociables.

In the third segment, tbe one whicb is particularlyconcerned with the flavor and appearance of theproduct, a large portion of the brand deciders arechildren. Their use of spearmint toothpaste is wellabove average. Stripe has done relatively well inthis segment. They are more ego-centered thanother segments, and tbeir life-style is outgoing butnot to the extent of the swingers. They will becalled The Sensory Segment.

The fourth segment, the price-oriented segment,shows a predominance of men. It tends to be aboveaverage in terms of toothpaste usage. People inthis segment see very few meaningful differences be-tween brands. They switch more frequently than

Benefit Segmentation: A Decision-oriented Research Tool 33

TABLE 1

TOOTHPASTE MARKET SEGMENT DESCRIPTION

Segment Name:

Principal benefit sought:

Demographic strengths:

Special behavioralcharacteristics:

Brands disproportionatelyfavored:

Personality characteristics:

Life-style characteristics:

The Sensory Segmetit

Flavor, product appearance

Children

Users of spearmintflavored toothpaste

Colgate, Stripe

High self-involvement

Hedonistic

The Sociables

Brightness of teeth

Teens, young people

Smokers

Macleans, Plus White,Ultra Brite

High sociability

Active

The Worriers

Decay prevention

Large families

Heavy users

Crest

High hypochon-drias is

Conservative

The IndependentSegment

Price

Hen

Heavy users

Brands on sale

Highautonomy

Value-oriented

people in other segments and tend to buy a brandon sale. In terms of personality, they are cognitiveand they are independent. They like to think forthemselves and make brand choices on the basis oftheir judgment. They will be called The Independ-ent Segment.

Marketing Implications ofBenefit Segmentation Studies

Both copy directions and media choices will showsharp differences depending upon which of thesesegments is chosen as the target—The Worriers, TheSociables, The Sensory Segment, or The IndependentSegment. For example, the tonality of the copywill be light if The Sociable Segment or The Sen-sory Segment is to be addressed. It will be moreserious if the copy is aimed at The Worriers. Andif The Independent Segment is selected, it will prob-ably be desirable to use rational, two-sided argu-ments. Of course, to talk to this group at all itwill be necessary to have either a price edge orsome kind of demonstrable product superiority.

The depth-of-sell reflected by the copy will alsovary, depending upon the segment which is of in-terest. It will be fairly intensive for The WorrierSegment and for The Independent Segment, butmuch more superficial and mood-oriented for TheSociable and Sensory Segments.

Likewise, the setting will vary. It will focus onthe product for The Sensory Group, on socially-oriented situations for The Sociable Group, andperhaps on demonstration or on competitive com-parisons for The Independent Group.

Media environments will also be tailored to thesegments chosen as targets. Those with serious en-vironments will be used for The Worrier and Inde-pendent Segments, and those with youthful, modernand active environments for The Sociable and theSensory Groups. For example, it might be logicalto use a larger proportion of television for The So-

ciable and Sensory Groups, while The Worriers andIndependents might have heavier print schedules.

The depth-of-sell needed will also be reflected inthe media choices. For The Worrier and RationalSegments longer commercials—perhaps 60-secondcommercials—would be indicated, while for theother two groups shorter commercials and higherfrequency would be desirable.

Of course, in media selection the facts that havebeen gathered about the demographic characteris-tics of the segment chosen as the target would alsobe taken into consideration.

The information in Table 1 also has packagingimplications. For example, it might be appropriateto have colorful packages for The Sensory Segment,perhaps aqua (to indicate fluoride) for The Wor-rier Group, and gleaming white for The SociableSegment because of their interest in bright whiteteeth.

It should be readily apparent that the kinds ofinformation normally obtained in the course of a.benefit segmentation study have a wide range ofmarketing implications. Sometimes they are use-ful in suggesting physical changes in a product.For example, one manufacturer discovered that hisproduct was well suited to the needs of his chosentarget with a single exception in the area of flavor.He was able to make a relatively inexpensive modi-fication in his product and thereby strengthen hismarket position.

The new product implications of benefit segmen-tation studies are equally apparent. Once a mar-keter understands the kinds of segments that existin his market, he is often able to see new productopportunities or particularly effective ways of posi-tioning the products emerging from his research anddevelopment operation.

Similarly, benefit segmentation information hasbeen found helpful in providing direction in the

34 Journal of Marketing, July, 1968

choice of compatible point-of-purchase materials andin the selection of the kinds of sales promotionswhich are most likely to be effective for any givenmarket target.

Generalizations fromBenefit Segmentation Studies

A number of generalizations are possible on thebasis of the major benefit segmentation studieswhich have been conducted thus far. For example,the following general rules of thumb have becomeapparent:

• It is easier to take advantage of market seg-ments that already exist than to attempt tocreate new ones. Some time ago the strategyof product differentiation was heavily empha-sized in marketing textbooks. Under this phi-losophy it was believed that a manufacturerwas more or less able to create new marketsegments at will by making his product some-what different from those of his competitors.Now it is generally recognized that fewer cost-ly errors will be made if money is first investedin consumer research aimed at determining thepresent contours of the market. Once thisknowledge is available, it is usually most effi-cient to tailor marketing strategies to existingconsumer-need patterns.

• No brand can expect to appeal to all consumers.The very act of attracting one segment mayautomatically alienate others. A corollary tothis principle is that any marketer who wishesto cover a market fully must offer consumersmore than a single brand. The flood of newbrands which have recently appeared on themarket is concrete recognition of this principle.

• A company's brands can sometimes cannibalizeeach other but need not necessarily do so. Itdepends on whether or not they are positionedagainst the same segment of the market. IvorySnow sharply reduced Ivory Flakes' share ofmarket, and the Ford Falcon cut deeply intothe sales of the standard size Ford because, ineach case, the products were competing in thesame segments. Later on, for the same com-panies, the Mustang was successfully introducedwith comparatively little damage to Ford; andthe success of Crest did not have a dispropor-tionately adverse effect on Gleem's marketposition because, in these cases, the segments towhich the products appealed were different.

• New and old products alike should be designedto fit exactly the needs of some segment of themarket. In other words, they should be aimedat people seeking a specific combination of bene-fits. It is a marketing truism that you sellpeople one at a time—that you have to getsomeone to buy your product before you get

anyone to buy it. A substantial group of peo-ple must be interested in your specific set ofbenefits before you can make progress in amarket. Yet, many products attempt to aimat two or more segments simultaneously. Asa result, they are not able to maximize theirappeal to any segment of the market, and theyrun the risk of ending up with a dangerouslyfuzzy brand image.

• Marketers who adopt a benefit segmentationstrategy have a distinct competitive edge. Ifa benefit segment can be located which is seek-ing exactly the kinds of satisfactions that onemarketer's brand can offer better than anyother brand, the marketer can almost certainlydominate the purchases of that segment. Fur-thermore, if his competitors are looking at themarket in terms of traditional types of seg-ments, they may not even be aware of the exist-ence of the benefit segment which he has chosenas his market target. If they are ignorant inthis sense, they will be at a loss to explain thesuccess of his brand. And it naturally followsthat if they do not understand the reasons forhis success, the kinds of people buying hisbrand, and the benefits they are obtaining fromit, his competitors will find it very difficult tosuccessfully attack the marketer's position.

• An understanding of the benefit segments whichexist within a market can be used to advantagewhen competitors introduce new products. Oncethe way in which consumers are positioningthe new product has been determined, the like-lihood that it will make major inroads intosegments of interest can be assessed, and a de-cision can be made on whether or not counter-actions of any kind are required. If the newproduct appears to be assuming an ambiguousposition, no money need be invested in defensivemeasures. However, if it appears that the newproduct is ideally suited to the needs of animportant segment of the market, the manufac-turer in question can introduce a new com-petitive product of his own, modify the physicalproperties of existing brands, change his adver-tising strategy, or take whatever steps appearappropriate.

Types of Segments Uncovered ThroughBenefit Segmentation Studies

It is difficult to generalize about the types ofsegments which are apt to be discovered in thecourse of a benefit segmentation study. To a largeextent, the segments which have been found bavebeen unique to the product categories being ana-lyzed. However, a few types of segments haveappeared in two or more private studies. Amongthem are the following:

Benefit Segmentation: A Decision-oriented Research Tool 35

The Status Seeker . . . a group which is verymuch concerned with theprestige of the brandspurchased.

The Swinger . . . a group which tries tobe modern and up todate in all of its activi-ties. Brand choices re-flect this orientation.

. . . a group which prefers tostick to large successfulcompanies and popularbrands.

. . . a group which looks forbenefits such as econo-my, value, durability,etc.

. . . a group which is espe-cially concerned withself-concept. Membersconsider themselves tohave a sense of humor,to be independent and/or honest.

. . . a group which is con-cerned primarily withsensory benefits.

Some of these segments appear among the cus-tomers of almost all products and services. How-ever, there is no guarantee that a majority of them

The Conservative

The Rational Man

The Inner-directed Man

The Hedonist

or, for that matter, any of them exist in any givenproduct category. Finding out whether they do and,if so, what should be done about them is the pur-pose of benefit segmentation research.

ConclusionThe benefit segmentation approach is of particular

interest because it never fails to provide fresh in-sight into markets. As was indicated in the tooth-paste example cited earlier, the marketing implica-tions of this analytical research tool are limitedonly by the imagination of the person using theinformation a segmentation study provides. Ineffect, when segmentation studies are conducted, anumber of smaller markets emerge instead of onelarge one. Moreover, each of these smaller marketscan be subjected to the same kinds of thoroughanalyses to which total markets have been subjectedin the past. The only difference—a crucial one—isthat the total market was a heterogeneous conglom-eration of sub-groups. The so-called average con-sumer existed only in the minds of some marketingpeople. When benefit segmentation is used, a num-ber of relatively homogeneous segments are uncov-ered. And, because they are homogeneous, descrip-tions of them in terms of averages are much moreappropriate and meaningful as marketing guides.

MARKETING MEMO

A difference between managers and scientists . . .

A basic controversy revolves around the question of problem vs. technique orienta-tion. In the extreme, managers are interested in problem solving regardless of tech-nique, and scientists are interested in sophistication of method regardless of applica-bility. Yet the question of pure vs. applied research is really one of degree ratherthan kind, and these designations are really related to long- and short-term con-siderations. However, overemphasis on pure research and technique developmentrather than problem solving has resulted in low yield from investments in manage-ment science. The percentage of recommendations that eventually afEect operations

has been miserably low.—^James E. Rosenzweig, Managers and

Management Scientists (Two Cul-tures) ," Business Horizons, Vol. 10(FaU, 1967), pp. 79-86, at pages 79,and 80.